insights
notes from the workshop.
Short pieces on how we think about AI systems, automation, and backend work - method and opinion, not press releases.
ai systems14 May 20265 min read
write the evaluation before you pick the model
The reason most AI features stall is not the model. It is that nobody can say whether a change made things better or worse. Start with the ruler, not the model.
readautomation28 March 20264 min read
automate the process you actually run, not the one on the diagram
The diagram is always cleaner than reality. The rules that matter usually live in one person’s head. Find those first, or your automation will be confidently wrong.
readbackend engineering9 February 20265 min read
the rewrite you probably don’t need
When a service gets slow, "rewrite it" is the reflex. It is also usually the most expensive way to fix a problem you have not measured yet.
readstart here
have a system that used to feel fast?
Tell us what it does and where it hurts. We’ll tell you honestly whether it needs sharpening or rebuilding - and what that would take.